Characterization of organic emissions from hazardous waste incineration processes under the new EPA draft Risk Burn Guidance: measurement issues

EPA's recently published draft Risk Burn Guidance recommends that hazardous waste combustion facilities complete a mass balance of the total organics (TOs) that may be emitted from the combustor. TOs, consisting of three dist inct fractions (volatile, semivolatile, and nonvolatile organic compounds) are determined using measurement techniques specific to the boiling point ( bp) range of each component. Preliminary proof-of-concept tests have been p erformed to gain further knowledge of the total chromatographable organics (TCO) and gravimetric analysis (GRAV) procedures used to determine the semi volatile and nonvolatile organic fractions. A commercially prepared Method 8270 semivolatile organic standard solution, made up of 112 compounds conta ining a variety of halogenated, oxygenated, nitrogenated, and sulfonated hy drocarbons, in addition to those hydrocarbons containing only carbon and hy drogen, was used to examine measurement biases of the TCO and GRAV methods. The results indicate that, at least for this particular standard, both met hods are accurate, exhibiting biases < +/- 10%. However, these TCO results may be deceiving. TCO measurement biases observed for select individual sem ivolatile organic compounds are well in excess of the composite bias observ ed. The composite bias observed may be more a result of a relatively equal number of negative and positive biases that negate each rather than a true indication of TCO measurement bias. Further methods evaluation work is requ ired to support this tenet. Additional tests were performed to verify that the semivolatile organic mass was not redundantly measured by the GRAV proc edure. Experiments were also performed to expand the mass of nonvolatile or ganic material that could be quantified using the TCO procedure. Expanding the TCO working bp range may be a viable alternative to the GRAV technique. It must be cautioned that these results are not fully comprehensive in nat ure and, as a result, applications of these results are limited. (C) 2000 E lsevier Science Ltd. All rights reserved.